SaaS and Cloud Outage Statistics: How Often Software Goes Down and What It Costs

Even a low-impact outage carried a median of $1.3M per hour in the 2024 survey, against $1.9M for high-impact events. The gap between the bars is smaller than most people expect, so a "minor" incident is still costly.

Respondents in the Americas reported a lower median ($1.4M) than Asia Pacific ($2.3M) or Europe ($1.7M). These are survey medians for large organizations, not a price list for your business.

U.S. Global 2000 companies averaged $256M a year in downtime costs, the highest of the three regions in the 2024 edition. No newer regional split has been published.

Lost revenue ($95M) is the biggest line, but regulatory fines ($51M) and ransomware payouts ($40M) show how closely downtime is tied to security and compliance problems.

Third-party or cloud provider failures (29%) sit just behind network failure (35%). Nearly a third of IT leaders blame an outside vendor, which is exactly where your contract terms do the work.

Amazon EC2 Region-Level SLA: service credits when monthly uptime falls below the 99.99% commitment

Source: Amazon Web Services, Amazon Compute Service Level Agreement. Download as image

Hidden business costs of downtime reported by Global 2000 executives (Splunk Hidden Costs of Downtime, 2026 and 2024 editions)

Source: Splunk (Cisco Newsroom), The $600 Billion Wake-Up Call: New Splunk Research Reveals Downtime Is a Systemic Business Crisis (The Hidden Costs of Downtime 2026); Splunk (Cisco Newsroom), Splunk Report Shows Downtime Costs Global 2000 Companies $400B Annually (The Hidden Costs of Downtime, with Oxford Economics). Download as image

How much does an hour of downtime cost?

For large organizations, the surveys agree on the order of magnitude. ITIC found that over 90% of mid-size and large enterprises lose more than $300,000 per hour of downtime, and 41% lose $1 million to over $5 million. New Relic's 2025 survey put the median cost of a high-impact outage at $2 million per hour. Splunk's 2026 edition works out to about $15,000 per minute on average.

Read the 2026 New Relic figure carefully. It reports an average of $1.85 million per hour for high-impact outages, while the 2025 number was a median. They measure different things, so the lower figure is not proof that outages got cheaper.

You are probably not a Global 2000 company, and your hourly number will be smaller. The method is the same, though. Add up revenue you can't take while the tool is down, wages for staff who sit idle, refunds or credits you owe your own customers, and the overtime to clean up afterward. Do that math once for each system your business can't work without. It tells you which contracts deserve a hard look.

How often do cloud and SaaS services go down?

New Relic's 2024 survey found a median of 77 hours of downtime a year across all impact levels. That is a little over three full days lost to outages, small and large.

The causes are a mix of things you control and things you don't. Network failure was cited by 35% of IT leaders, third-party or cloud provider failures by 29%, someone changing the environment by 28%, and software deployments by 27%. Uptime Institute counted IT and networking issues behind 23% of impactful outages in 2024.

One more number should worry any team that sells online. In Splunk's 2026 research, 47% of tech executives said customers are often or very often the first to spot an outage. If your vendor's status page is green while your customers are emailing you, you need a contract that defines downtime by what you can measure as well as what the vendor reports.

Why won't a vendor's SLA credit cover my losses?

Because a service credit is a discount on your bill, not a payment for your loss. Amazon's EC2 regional SLA is a good example. If monthly uptime drops below 99.99% but stays at or above 99.0%, you get 10% of that month's EC2 bill for the affected region. Between 99.0% and 95.0% it rises to 30%. Only below 95.0% do you get the full month back.

Say you pay a few thousand dollars a month for hosting and lose a full business day. The credit is a slice of that bill. Your lost sales, idle staff and angry customers are not part of the formula.

Many SaaS contracts then add two more layers. The SLA often says credits are your "sole and exclusive remedy" for downtime, and the liability cap limits everything else, commonly to fees paid over a short window. Put those together and a vendor outage that costs you real money may entitle you to almost nothing. That is why the SLA, the remedy wording and the liability cap need to be read as one package.

What does downtime cost beyond lost revenue?

Splunk's 2026 edition breaks direct costs down per Global 2000 organization: $95M in lost revenue, $51M in regulatory fines and $40M in ransomware payouts each year. The last two lines are security and compliance costs, not technical ones. In the 2024 edition, 56% of downtime incidents traced back to security incidents, not hardware or bugs.

There are slower costs as well. Companies saw an average stock price drop of 3.4% after a downtime event in the 2026 research. The 2024 edition found lost revenue took 75 days to recover, and 74% of tech executives reported delays in getting products to market.

For a smaller business the equivalent is a lost client, a missed launch or a refund wave. Uptime Institute's 2025 survey found 57% of operators said their most recent major outage cost more than $100,000. A contract that only promises service credits does nothing for any of this.

Are outages getting more expensive?

On the broad measure, yes. Splunk's estimate for Global 2000 companies rose to $600B a year in 2026, a 50% jump in two years. Lost revenue per organization nearly doubled over the same stretch.

New Relic's annualized figure moved the other way, from $76M in 2025 to $73.5M in 2026, though the report doesn't say whether the newer number is a median or an average. The fair reading is that outage costs are high and not falling in any clear way. Plan as if the next one will cost at least as much as the last.

What should a business or freelancer do about outage risk?

You can't stop AWS or your payroll tool from going down. You can decide in advance who carries the cost and how fast you get your data back. Start with the tools that would stop work entirely if they failed.

• Map your critical vendors. List every SaaS and cloud service your revenue depends on and estimate what an hour without each one costs you.
• Read the SLA, not the marketing page. Check the uptime percentage, how downtime is measured, what counts as excluded maintenance, and whether you must file a claim to get a credit.
• Look for the "sole remedy" line. If credits are your only remedy, push for a right to terminate after repeated failures and a carve-out from the liability cap for data loss.
• Match what you promise downstream. Freelancers and agencies who host client sites or apps should never promise more uptime to a client than their own provider promises them.
• Get your data out on your terms. Make sure you can export your data at any time, in a usable format, while the contract is still running.
• Keep the incident trail. Save screenshots, timestamps and support tickets during an outage, because many SLAs require you to claim credits within a short window.

Sources and method

Every number on this page comes from the sources below, each opened and read on or before October 18, 2026. A second, independent review checked every figure against its source and corrected anything that did not match. Figures that could not be confirmed were left out.

1. Information Technology Intelligence Consulting (ITIC), ITIC 2024 Hourly Cost of Downtime Report (September 3, 2024)
2. Information Technology Intelligence Consulting (ITIC), ITIC 2024 Hourly Cost of Downtime Part 2 (September 10, 2024)
3. Splunk (Cisco Newsroom), The $600 Billion Wake-Up Call: New Splunk Research Reveals Downtime Is a Systemic Business Crisis (The Hidden Costs of Downtime 2026) (May 19, 2026)
4. Splunk (Cisco Newsroom), Splunk Report Shows Downtime Costs Global 2000 Companies $400B Annually (The Hidden Costs of Downtime, with Oxford Economics) (June 11, 2024 (2024 edition; superseded by the 2026 edition of May 19, 2026))
5. Uptime Institute, Uptime Announces Annual Outage Analysis Report 2026 (8th Annual Outage Analysis) (May 13, 2026)
6. Uptime Institute, Uptime Announces Annual Outage Analysis Report 2025 (7th Annual Outage Analysis) (May 6, 2025)
7. New Relic, 2024 Observability Forecast: Outages, downtime and cost (2024 edition (superseded by the 2025 and 2026 editions; cited only for measures the 2026 edition does not report))
8. New Relic, New Relic 2025 Observability Forecast press release (September 17, 2025)
9. New Relic, New Relic 2026 Observability Forecast Report (2026 edition (PDF posted September 2026, © 2026))
10. Amazon Web Services, Amazon Compute Service Level Agreement (Last updated May 25, 2022)